Facebook's datacenter photo, credit: Robert Scoble at Flickr
Facebook — Bots Once Again Appear
Courtesy of Karl Denninger
This has come up before, but the charge just got much more serious.
Any business that advertises on Facebook wants eyes looking at its ads and then fingers clicking on them. Facebook gets paid based on how many clicks that ad receives – effectively, on how many users it can send to a particular brand.
On Monday came an explosive claim that could give pause to brands trying to figure out if advertising works on Facebook. A Long Island start-up company said it was pulling its ads from the social network because it discovered that its ad clicks were far more likely to be coming from Web robots – or bots, as they are known — than human Facebook users.
Um, for Facebook's sake, I hope not.
It's not that difficult to detect this sort of thing either. See, Facebook relies on Javascript. A lot. In fact, turn it off and then see what Facebook looks like. Be prepared to be amused; all the "hover over" things and similar stop working.
So what did the firm do? It ran its own analytics and found that most of the so-called "clicks" that it was paying for were coming from something on the other end that didn't have javascript turned on. While it's possible that some of the users have it off or are "adblocking" in some way, it makes absolutely no sense that 80% of the people allegedly "clicking" an ad are doing so in this fashion.
In other words, most of those "clicks" were almost-certainly a robot.
Facebook isn't necessarily the one doing it; indeed the company was careful to say that it didn't know who was doing it. In fact, it's entirely possible for someone to "target" an advertiser in this fashion, and it's done from time to time.
But ad networks such as Google's spend quite a bit of time and effort to identify and put a stop to this sort of misbehavior. You can't prevent the robot from "clicking" the ad but you can, as an ad network, do your level best to identify what "clicks" are bogus and not charge the advertiser for them.
In terms of damage to an ad network's revenue model (which is what Facebook is, really, when you get down to it) it doesn't make much difference who is doing the scamming. What matters is that if advertisers are being charged for delivery of ads to robots instead of people then they're being charged for something that didn't happen (exposure of their ad to a pair of eyeballs.)
Watch this one carefully folks — defections of advertisers over problems like this have the potential to cause real trouble.
